arvol

• using webized prolog which basically means that atoms can be IRIs
• besides top-down reasoning with conclusion :- premise rules, it also supports bottom-up reasoning with conclusion :+ premise rules
• variables are interpreted universally except for conclusion :+ premise conclusion-only variables which are interpreted existentially
• bottom-up reasoning can use stable(n) to fail if the deductive closure at level n is not yet stable
• bottom-up reasoning steps are performed as step((conclusion :+ premise), premise_inst, conclusion_inst)
• queries are posed as true :+ premise and answered as answer(premise_inst)
• inference fuses are defined as false :+ premise and blown as fuse(premise_inst) with return code 2

rationale for bottom-up reasoning

• conclusion can be a conjunction
• conclusion can be false to blow an inference fuse
• conclusion can be true to pose a query
• conclusion can not be any other built-in
• conclusion-only variables are existentials
• performing bottom-up proof steps step/3
• avoiding loops that could occur with top-down reasoning

test

• using eye

git clone https://github.com/eyereasoner/arvol
cd arvol/etc
./test

• using trealla

git clone https://github.com/eyereasoner/arvol
cd arvol/alt
./test

background

• personal notes by Tim Berners-Lee: Design Issues
• book of Markus Triska: The Power of Prolog